Optical interconnects based on high-contrast all-dielectric nano-post arrays

Area of Science:

• Optoelectronics
• Nanophotonics
• Integrated Neural Networks

Background:

• Optical interconnects are crucial for high-speed data transfer in optoelectronic integrated neural networks.
• Existing solutions like zone plates face limitations in focal length and efficiency for compact designs.

Purpose of the Study:

• To present a novel, compact solution for optical interconnects using high-contrast all-dielectric nano-post arrays.
• To demonstrate the potential of these nano-post arrays for efficient light manipulation in neural network applications.

Main Methods:

• Fabrication of nano-post arrays using amorphous silicon, chosen for its high refractive index and near-infrared transmittance.
• Tuning individual post radii to precisely control phase delays for light deflection and convergence.
• Numerical simulations to calculate connection efficiencies and compare performance against zone plate interconnects.

Main Results:

• The proposed nano-post arrays enable precise control over light's phase profile, achieving desired deflection and convergence.
• Numerical simulations indicate superior performance of nano-post arrays over zone plates, particularly for short focal length requirements.
• High connection efficiencies are achievable with the novel nano-post array design.

Conclusions:

• Amorphous silicon nano-post arrays provide a compact and efficient solution for optical interconnects in optoelectronic neural networks.
• The demonstrated technology offers significant advantages over traditional zone plates for specific applications demanding high efficiency and short focal lengths.